The variability in the prognosis of individuals with hepatocellular carcinoma (HCC) suggests that HCC may comprise several distinct biological phenotypes. These phenotypes may result from activation of different oncogenic pathways during tumorigenesis and/or from a different cell of origin. We have address whether the transcriptional characteristics of HCC can provide insight into the cellular origin of the tumor. We integrated gene expression data from rat fetal hepatoblasts and adult hepatocytes with HCC from human and mouse models. Individuals with HCC who shared a gene expression pattern with fetal hepatoblasts had a poor prognosis. The gene expression program that distinguished this subtype from other types of HCC included markers of hepatic oval cells, suggesting that HCC of this subtype may arise from hepatic progenitor cells. Analyses of gene networks showed that activation of AP-1 transcription factors in this newly identified HCC subtype might have key roles in tumor development.In this work we have shown that by applying two independent gene expression signatures, we were able to divide individuals with HCC into three subgroups characterized by statistically significant differences in clinical outcome. These findings support the notion that multiple molecular pathways dictate the development and different clinical outcomes of HCC. Our finding also indicates that the molecular features of HCC such as prognostic gene expression signatures are present at the time of diagnosis. Therefore, the use of gene expression profiling promises to improve molecular classification and prediction of outcomes in HCC. Furthermore, molecular stratification of individuals with HCC into homogeneous subgroups may provide opportunities for the development of new treatment modalities.